We will be using twitter to continue the conversation about computational linguistics outside of class.
We will mark all tweets from this class with the hashtag #cl472. In addition, we will develop a classification of tweet types we want to use (and associated hashtags) in the first week of class. This will allow us to use the corpus as labeled training data for a classification task.
In class on 3/28, we decided on a set of labels that reflect both dialogue act (question/response/statement/exclamation) and emotion: #curious, #help, #fyi, #wtf, #hth, #argh, #yay, #lol. To the extent possible, all tweets should be labeled with one of these. If none of them fit, the tweet belongs to the "other" category. (We'll assume all unlabeled tweets are "other".)
Note that twitter is not intended to replace the class dicussion board. The discussion board will still be used to discuss assignments and well as course content. We do expect to see cross-references (links) between twitter and the discussion board!
Towards the end of the quarter, we'll reflect on to what extent this exercise was useful, and how.